ArcGIS Online has a new function that enable you to subset data in an active hosted feature class called a Hosted View.

While the hosted view is great, there are valid reasons why you may not want to leverage this functionality.

The method described here is to archive data from the active Feature Layer into a separate matching archive layer.  Some benefits of using this strategy include

  1. Reducing the amount of data in the primary hosted layer that you are leveraging in your day to day business
  2. Improving performance of the primary layer depending on how much data you have in it
  3. Reducing credit usage if you archive to an external system, such as uniquely named FileGeodatabases which are stored on your file system or other file storage system.

This specific example shows a very simple workspace moving data from an active layer to an archive layer.  This particular workspace works beautifully on FME Server or FME Cloud.  Clearly it’s a little simple for full production use, but it highlights an easy way to get data from one system to another.

Creator

Well, you need a Creator, because that is simply how FME works to kick things off.

FeatureReader

This feature reader, as setup will read all of the features in the defined ArcGIS Online hosted feature layer.

FeatureWriter

The FeatureWriter transformer connects to the defined layer and writes all of the data read from the ArcGIS Online hosted layer to the equivalent archive layer.

FeatureWriter

This particular FeatureWriter has been setup for Delete mode and will delete everything in the ArcGIS Online Hosted Feature Layer.

Further work and refinement needs to be done to “productionize” this system.  For example, the use of Terminator’s in case there is an error as well as Null writers to collect statistics for reporting purposes.  Parameterizing the system will also enable you to extend it further.

We have used this general technique in a number of cases, typically where there is routine data collection over the same area and the client only wanted to keep the most recent data in their active maps.  In that case we removed data from the active layer and moved it to an archive layer if they ever wanted to go back for analysis purposes.

The other cases are for very high volume data processing, in one case the mobile units are collecting and storing approximately 32,000 locations per hour.  On a nightly schedule, the data is processed to support a variety of reporting requirements and then archived to an external environment in order to control the amount of data in the active system.

Happy FME’ing